Twenty-three studies (at least 4,238 participants; exact number unclear) were included in the review.
All of the studies were double-blind except for one, which was single-blind. Six of the studies had adequate allocation concealment; for the rest, allocation concealment was unclear.
ICS at all doses and regardless of device were associated with a significantly greater incidence of oral candidiasis (OR 3.6, p<0.001; 20 studies), dysphonia (OR 5.2, p<0.001; 15 studies), and pharyngitis (OR 2.2, p<0.023; 16 studies) than placebo.
Compared with placebo and regardless of dose or device, fluticasone propionate had the greatest risk of oral candidiasis (OR 5.41, p<0.001) while budesonide had the greatest risk of both dysphonia (OR 11.45, p=0.02) and pharyngitis (OR 5.09, p=0.04).
The MDI device was associated with a 5-fold greater risk of oral candidiasis compared with placebo (OR 5.40), whereas the DPI device only showed a 3-fold greater risk of oral candidiasis compared with placebo (OR 3.24). Trends for dysphonia were similar (OR 5.68 and OR 3.74 for MDI and DPI, respectively). The risk for dysphonia was increased 2-fold when using either device.